News Archive
Pick a date to read that day's stories.
Days with a blue dot have stories.
- 01 of 10ANTHROPIC
Claude Fable 5 quietly returns through the Android app
A popular story disappeared from an app but briefly returned. This has worried people, and new access rules may follow.
OUR TAKEBack online, and somehow that is more worrying.
Read more → - 02 of 10POLICY
Macron asks Trump to share America's best AI with allies
Macron asked Trump to share advanced US computing with allies. Trump did not agree, increasing Europe's worry about being left behind.
OUR TAKEFrance would like the iPhone of AI, please, and a receipt.
Read more → - 03 of 10SAFETY
Amazon says big-company AI oversight is broken
Amazon says human oversight of many AI choices fails. People make mistakes, so companies must find new ways to check AI behavior.
OUR TAKEThe job title AI supervisor already needs a supervisor.
Read more → - 04 of 10POLICY
Europe wants in on Anthropic's AI, Washington says no
Macron and Trump clashed over whether European countries should get access to Anthropic's most powerful models. Washington wants to keep them on US soil. Brussels is now openly talking about building its own.
OUR TAKENothing motivates a continent like being told it cannot have something.
Read more → - 05 of 10POWER
US regulator orders six grid operators to plug AI in faster
US energy regulators want six grid operators to quickly explain how they'll connect AI power needs. This move could mean regular people end up paying more for electricity.
OUR TAKEYour next electricity bill may include a small thank-you from ChatGPT.
Read more → - 06 of 10POLICY
G7 leaders spend their summit arguing about AI
Leaders at a major summit discussed artificial intelligence. Top tech company bosses also attended, showing its global importance.
OUR TAKEThe G7 has discovered there is no future not run on someone else's servers.
Read more → - 07 of 10MODELS
GPT-5.6 is quietly being tested on paying ChatGPT users
Some ChatGPT Pro subscribers are getting noticeably sharper answers, with coding and reasoning visibly improved. OpenAI has said nothing, which usually means an official launch is days away. The pattern matches every previous big release.
OUR TAKEIf your ChatGPT got cleverer this weekend, it is not your imagination.
Read more → - 08 of 10EU
Estonia formally designs legal IDs for AI agents
Estonia will give bots legal ID numbers. This lets people see what each bot can do.
OUR TAKEEstonia: small country, very large reputation for going first.
Read more → - 09 of 10ANTHROPIC
Anthropic adds a second pair of eyes to Claude Code
Anthropic's new tool helps teams check AI-written code together. Reviewers can comment and approve changes, making offices more comfortable with AI.
OUR TAKEThe AI writes it, humans still sign it, for now.
Read more → - 10 of 10RESEARCH
Study: people refuse to act on AI lie detectors
People ignore computers that detect lies because they dislike accusing others. Even when the computer is right, people dismiss its findings. Accuracy doesn't always lead to usefulness.
OUR TAKEWe built the lie detector, then politely looked the other way.
Read more →